A new study of neural prostheses in monkeys suggests that learning to control a robotic arm with the power of thought may happen more naturally than scientists had expected. Jose Carmena and Karunesh Ganguly at the University of California, Berkeley (UCB), found that the animals create a mental map of the device, much as we do when learning to swim or swing a tennis racket.
A number of labs have already shown that monkeys–and in a few cases, humans–with electrodes implanted into their brains can learn to control a computer cursor or robotic arm. To train the subjects, scientists first record the activity in a group of neurons as the monkey moves its real arm (or in the case of a paralyzed human, as the person imagines moving his or her arm). Researchers then analyze the neural activity to develop a decoding algorithm that can translate the pattern of brain-cell firing into an action–say, moving a cursor to a certain point on a computer screen.
In the new study, published today in the journal PLoS Biology, monkeys learned to precisely control a computer cursor over a few days. As the animals became proficient at the task, the researchers identified a specific pattern of neural activity in the brain associated with the movement. “The profound part of our study is that this is all happening with something that is not part of one’s own body. We have demonstrated that the brain is able to form a motor memory to control a disembodied device in a way that mirrors how it controls its own body. That has never been shown before,” said Carmena in a press release from UCB.
The research also suggests that optimizing the decoders may not be as important as expected. A few weeks after the monkeys learned to control the arm with the original decoder, Carmena and Ganguly introduced a new one, indicated by a different colored cursor. According to the release:
As the monkeys were mastering the second decoder, the researchers would suddenly switch back to the original decoder and saw that the monkeys could immediately perform the task without missing a beat. The ability to switch back and forth between the two decoders shows a level of neural plasticity never before associated with the control of a prosthetic device. “This is a study that says that maybe one day, we can really think of the ultimate neuroprosthetic device that humans can use to perform many different tasks in a more natural way,” said Carmena.
DeepMind’s cofounder: Generative AI is just a phase. What’s next is interactive AI.
“This is a profound moment in the history of technology,” says Mustafa Suleyman.
What to know about this autumn’s covid vaccines
New variants will pose a challenge, but early signs suggest the shots will still boost antibody responses.
Human-plus-AI solutions mitigate security threats
With the right human oversight, emerging technologies like artificial intelligence can help keep business and customer data secure
Next slide, please: A brief history of the corporate presentation
From million-dollar slide shows to Steve Jobs’s introduction of the iPhone, a bit of show business never hurt plain old business.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.